The AI Readiness Gap

There is a question that sits underneath every retail AI summit, and it is almost never the one on the agenda. The agenda asks which model, which vendor, which use case to back first. They are fair questions. But I think they are all second questions. Before you ask what AI can do for your organisation, it is worth asking whether your organisation is ready to be changed by it.

Do you have a Readiness Gap?

I know how that sounds. Readiness is the least glamorous word in the room. It has none of the shine of a live demo or a pilot that lights up on stage. And yet the evidence for taking it seriously is now hard to look away from. MIT's State of AI in Business 2025 study, drawn from around three hundred public deployments and more than a hundred and fifty executive interviews, found that ninety-five per cent of enterprise generative AI pilots delivered no measurable return. Not because the technology failed. Because the organisation around it was never built to absorb what the technology could do. The pattern has held. By January 2026, Gartner reported that at least half of generative AI projects had been abandoned after proof of concept, and the causes it names, poor data quality, inadequate risk controls, escalating costs and unclear business value, are organisational failures, not model failures.

The Readiness Gap is the distance between the capability you can buy and your organisation's ability to actually carry it.

That gap is worth naming. The retail floor is not littered with bad AI. It is littered with good AI that nobody adopted, sitting in a folder somewhere.

Here is what makes this moment different, and a little unforgiving. Earlier waves of retail technology you could mostly bolt on. A new POS, a loyalty platform, an app. AI is not like that. It asks the organisation to change how it decides, who it trusts with those decisions, and how the work actually gets done. You are not installing a tool. You are asking people to work differently, and quite often to trust a machine with judgement they used to own. That is a human question long before it is a technical one.

I say this with the humility of someone who has sat in the room when it went wrong. I have watched a pilot dazzle a leadership team, watched the nods go around the table, and watched precisely nothing happen in the six months that followed. Not through anyone's incompetence. The purpose was fuzzy, so nobody could point the thing at anything that mattered. The people were not ready to trust it, so they quietly routed around it. And there was no repeatable path from the clever pilot to the embedded reality. The technology was never the problem. We were.

Three quiet questions

So what does readiness actually mean, once you strip away the change-management jargon? I would offer that it comes down to three quiet questions. Is your purpose clear enough that you can aim AI at something that matters, rather than at whatever demos well? Are your people ready to trust it, and ready to be trusted with it? And is there a repeatable way to get from a promising pilot to something the organisation uses without being told to?

Those three questions are the reason we ended up building three models at ThinkUncommon, one that acts on purpose, one that acts on people, and one that acts on the work itself. Together they are the Readiness Stack, and the cards below will take you to them. The bigger point is that readiness is not a single foundation slide you nod past on the way to the good stuff. It has structure, it can be assessed, and it is far cheaper to build it before the money goes out the door.

None of this is an argument against AI. Quite the opposite. AI is the most significant medium retail has been handed in a generation, and the retailers who use it well will pull ahead. But a medium is only as good as what an organisation can do with it. The tool is not the strategy. Readiness is the thing that turns the tool into a result.

Which is why I would push back on how we frame an AI agenda in the first place. Readiness is not the dull preamble before the exciting technology content. It is the thing that decides whether any of the exciting technology content ever pays for itself. A summit that covers only the tools is, with respect, doing about half the job.

The retailers who win the AI decade will not, I suspect, be the ones who moved first. They will be the ones who were ready to be changed. That is harder to put on a slide. But it is where the results will come from.